Scaling the training of Large Language Model (LLM) agents typically requires a large collection of executable environments, whose construction, maintenance, …
机构:美团
来源:arXiv 2610.27847 | AI4Papers 论文推荐平台